The appeal of adaptive trial designs is undeniable. Flexible models for clinical research give sponsors the ability to shift course midstream – dropping arms, adding patients, focusing on subpopulations, and making other adaptations based on early results.

These adjustments can cut costs and improve operational efficiency while increasing the chance of better research outcomes. It is the ‘value trifecta’ that sponsors have been chasing for years.

But for all the hype about adaptive trial designs, a lot of clinical teams still struggle with how to make them work.

If sponsors want to take advantage of the benefits of an adaptive trial design, they need to build it into their clinical trial plan – before the study begins. That requires extra time and effort up front, along with the right talent, and robust modeling tools to determine which adaptive models will meet their needs.

We understand that adaptive designs are complicated and that it takes expertise and guidance to get them right. So our Pipeline Architect team created a video on how to incorporate adaptive trial designs into clinical trial plans. This 11-minute course lays out four basic adaptive models and offers advice how to build and evaluate trials using each of them.

Here are some of the highlights.

Four adaptive trial designs

In the video, we discuss four common adaptations that can help sponsors cut time and costs, reduce patient exposure, and add efficiency to their R&D. In each model, the trial plan defines key milestones in the research – before all patients are recruited – where data will be analyzed to determine whether a shift in direction is warranted.

It is important to be very specific about the timing and conditions for these analysis milestones and to include the specific indicators that support the adaptation.

Sample size re-estimation

Clinical trials almost always specify a sample size or number of patients to recruit. In most cases, these sample sizes are determined based on an estimate of the minimum number of patients likely to generate meaningful results.

However, the sample size is calculated based on assumptions that may not be accurate. Using sample size re-estimation, sponsors can assess patient data at a defined milestone to determine whether an adjustment to the trial size is needed. For example, a trial may anticipate recruiting 160 patients, but using sample size re-assessment, the sponsor will analyze pooled data from the first 40 patients to determine if more patients will be needed to deliver the desired results.

Such adaptations can cut significant time and cost from the plan on average and reduce risk of trial failure -- though if the data does not support adaptation, recruiting additional patients will add time and cost to the project.

Benefits

• Sample-size re-estimation allows sponsors to right-size the number of patients, reducing costs and accelerating time to market.
• It allows sponsors to avoid negative studies, if data indicates the compound is efficacious but a larger sample size is required.
• This model can be appealing to financial decision-makers who see an opportunity to potentially lower costs and avoid risk.

Interim analysis with stopping for futility or efficacy

In this adaptive model, sponsors conduct an interim analysis at a pre-defined milestone to assess how the trial is going. This may uncover a surprisingly positive treatment effect that suggests the trial can be concluded based on current results. Or, it may show that the lack of treatment effect (or unexpected safety concerns) means the trial should be stopped and no further patients enrolled.

Benefits

• Whether the data is positive or negative, sponsors may come to important conclusions sooner in the research process to support current and future investment decisions.
• If it is clear that the endpoints can’t be achieved, data may justify ending the trial and not exposing more patients to a treatment that does not show benefit.
• If the treatment effect is overwhelmingly positive, sponsors may be able to accelerate time to market.

Dropping an arm

In this model, sponsors launch multiple arms of a trial knowing that one or more will be cut. This is a popular option for companies that want to test broader dosing ranges or additional sub-populations to determine which group might experience the most positive results.

Benefits

• Planning to drop arms gives sponsors the flexibility to test a broader range of dosing strategies, which increases likelihood of development success.
• It’s an appealing model for investors and researchers, because more up-front data reduces risk, while guaranteeing a reduction in costs when sub-efficacious arms are dropped.
• It lowers patient risks by ending poorly performing arms sooner, reducing patient exposure to less effective dosages.

Adjusting the study population

Sponsors who believe that a treatment will more positively affect certain subpopulations can validate their theories by using an adaptive design to adjust the population mid-trial (e.g. population enrichment). Analyzing data at an interim milestone lets sponsors assess whether certain populations are experiencing a higher treatment effect, then adjust their inclusion/exclusion criteria for the remaining recruits.

The key to this model is that the subpopulations and data-driven decisions must be predefined in the study plan with a strong rationale to maintain scientific integrity. Regulators will not support sponsors going on a hunting expedition midway through the trial.

Benefits

• Population adjustment allows sponsors to recruit more rapidly from a broader population, with a backup plan to focus on a smaller, more responsive population only if warranted.
• It reduces the likelihood of a negative trial by enhancing the population with the greatest treatment response to drug.
• It lowers patient risks by recruiting fewer populations that aren’t likely to respond or which have a higher likelihood of adverse events.

Each of these models can bring benefits, but only if sponsors give their clinical trial designers the additional time and support to do the design research and analysis up-front. This can be a challenge, particularly when teams face pressure to complete planning as quickly as possible so recruiting can begin.

Trial designers can accelerate this process by using sophisticated modeling tools, like IQVIA’s Pipeline Architect. The technology helps identify cases where adaptations can deliver significant value, model the ideal timing and patient scenarios, and provide detailed evidence needed to make a case to decision-makers as to the benefit of adaptive designs versus standard fixed designs.

View the 11-minute video explaining some common adaptive designs in action.